e-Infrastructure and e-Services for Developing Countries. 8th International Conference, AFRICOMM 2016, Ouagadougou, Burkina Faso, December 6-7, 2016, Proceedings

Research Article

Multi-diffusion Degree Centrality Measure to Maximize the Influence Spread in the Multilayer Social Networks

  • @INPROCEEDINGS{10.1007/978-3-319-66742-3_6,
        author={Ibrahima Gaye and Gervais Mendy and Samuel Ouya and Idy Diop and Diaraf Seck},
        title={Multi-diffusion Degree Centrality Measure to Maximize the Influence Spread in the Multilayer Social Networks},
        proceedings={e-Infrastructure and e-Services for Developing Countries. 8th International Conference, AFRICOMM 2016, Ouagadougou, Burkina Faso, December 6-7, 2016, Proceedings},
        proceedings_a={AFRICOMM},
        year={2017},
        month={10},
        keywords={Centrality measure Diffusion probability Influence maximization Mapping matrix Multilayer social network},
        doi={10.1007/978-3-319-66742-3_6}
    }
    
  • Ibrahima Gaye
    Gervais Mendy
    Samuel Ouya
    Idy Diop
    Diaraf Seck
    Year: 2017
    Multi-diffusion Degree Centrality Measure to Maximize the Influence Spread in the Multilayer Social Networks
    AFRICOMM
    Springer
    DOI: 10.1007/978-3-319-66742-3_6
Ibrahima Gaye,*, Gervais Mendy,*, Samuel Ouya,*, Idy Diop,*, Diaraf Seck,*
    *Contact email: gaye.ibrahima@esp.sn, gervais.mendy@ucad.edu.sn, samuel.ouya@gmail.com, idy.diop@esp.sn, diaraf.seck@ucad.edu.sn

    Abstract

    In this work, we study the influence maximization in multilayer social networks. This problem is to find a set of persons, called seeds, that maximizes the information spread in a multilayer social network. In our works, we focus in the determination of the seeds by proposing a centrality measure called - (denoted by ) based on model. We consider the persons as the most influential. This centrality measure uses firstly, the diffusion probability for each person in each layer. Secondly, it uses the contribution of the first neighbors in the diffusion process. To show the performance of our approach, we compare it with the existing heuristics like . With software and , we show that - is more performant than the benchmark heuristic.